IEEE’s Post

Join me to explore the incredible world of Data Mining and Neural Networks. In this video, you'll discover: 📊 Data Mining: What it is and why it's crucial in today's data-driven world. 🧠 What is the Neural Network: An introduction to neural networks and their fundamental concepts. 🔄 How the Neural Network Works: A deep dive into the mechanics behind neural networks. 🔍 Neural Network Types: A look at various types of neural networks and their distinct functions. 💡 Applications: Practical applications and real-world examples of neural networks in action. Whether you're new to the field or looking to expand your knowledge, this presentation covers it all. Check it out! #DataScience #NeuralNetworks #ArtificialIntelligence #MachineLearning #DeepLearning #EducationalVideo #LinkedInLearning

Support for graph neural networks in TensorFlow! ▶ TensorFlow GNN 1.0 (TF-GNN) ▶ pip install tensorflow_gnn 🔸 What are some examples of graphs? Lots of items in the real world can be defined in terms of their connections to other things. This could be social network data, knowledge graphs, molecules, even images represented as graphs (TLDR; graphs are everywhere) 🔸 How do we represent graphs and what makes them unique? Items of a graph are represented by nodes, and the connections between them are represented by edges. Graph structured data is inherently different than other types of sequential or tabular data representations often used in ML 🔸 What is a graph neural network? A GNN is a type of neural network designed to process and analyze graph structured data and its attributes (nodes, edges, global context), without changing the connectivity of the input graph. GNNs are able to make predictions based on relationships between elements in a graph TensorFlow Blog Post: https://lnkd.in/gkeYMDpw Intro to graph neural networks: https://lnkd.in/gRxtzB_v #deeplearning #graph #GNN

A long wait but this paper is finally accepted to IEEE Transactions on Neural Networks and Learning Systems #TNNLS From Bi-Level to One-Level: A Framework for Structural Attacks to Graph Anomaly Detection https://lnkd.in/ghAxZdMp

Hello colleagues, I recently finished the first research article on the topic - Security Issues of Deep Learning Neural Networks Machine Learning from Backdoor Attacks. The article was written in Russian, I’ll add its translation later. The research was published in the scientific journal “Zaŝita informacii. Inside"

Hi all, I have successfully completed the "Convolutional Neural Networks" course offered by Great Learning Academy. This course provided a solid foundation in CNNs, including essential concepts like architecture, applications, and batch normalization techniques🌟 #GreatLearningAcademy #greatlearning #glacertificate

My paper "Supercharging BKT with Multidimensional Generalizable IRT and Skill Discovery" just got published in the Journal of Educational Data Mining. It creates two novel models that can discover the skills required to solve questions in tutoring systems, capture variations in question difficulties, generalize to new students, and provide interpretable parameter estimates (unlike black box neural network models). Link:

🎉 I am pleased to announce the publication of my first paper on quantifying the uncertainty of soil type maps predicted with neural networks! 📚 This paper, published in the journal Science of the Total Environment (STOTEN), investigates the Last-Layer Laplace Approximation method in the context of soil type maps and an associated uncertainty specification incorporating model uncertainty. I am grateful for all support along the way. Check it out: https://lnkd.in/etCE9KKx #FirstPaper #Research #AcademicTwitter #ProudMoment

Use of Deep Learning in De-mining 🤔 The program is designed to distinguish amongst nine distinct materials with an exceptional validation accuracy of 93.5%. This high precision enables the differentiation between anti-personnel mines (APM) and small vertical paper pins (0.2 g) in various environments, including mineralized soil and non-mineralized environments like air or sand. By leveraging deep neural networks, the program offers improved accuracy and effectiveness compared to traditional classifiers like K-Nearest Neighbors and Support Vector Machines. One of the notable challenges addressed by the A.I. Model is accurately classifying APM buried in sand, where false negatives pose a significant risk. While the DL model exhibits a 5.2% false negative rate in these scenarios, it boasts a high confidence rate of 94.8% in correctly classifying the target. Coherence Coherence Recent (Complete Publication) : https://lnkd.in/gbCwGqgf

Our paper which experimented training a general neural network to explain any CNN has been accepted in the prestigious Knowledge Based Systems Journal [Impact factor 7.2]. Our work trained a GAN to produce attribution maps to explain CNNs supervised by GradCAM maps. We found that the attribution maps deviated from GradCAM and these deviations were synonymous to explanations generated by multiple post-hoc explainers. We also show the importance of both initial and deeper layers of the CNN for making predictions through mechanistic interpretations of the GAN. I thank my supervisors Sikha O K and Raul Benitez Iglesias for their guidance and support throughout my research journey. Link to Paper: https://lnkd.in/g9RsCuNy Link to code: https://lnkd.in/g_3kK-sY

I'm thrilled to announce that I have completed the fourth course of the Deep Learning Specialization offered by DeepLearning.AI on Coursera🌟.It is about Convolutional Neural Networks (CNN). I have learnt about: - The Convolution Operation. - The Convolutional layer/ The Pooling layer / The FC layer. - YOLO library. - Object Detection and Segmentation. - TensorFlow. For more to come ان شاء الله #Achievement #Certificate #deeplearning #deeplearningai